Section: Overall Objectives
An important scientific step was reached by proposing a first formal, as well as experimental, combination of two central families of exploration and learning algorithms, which were so far studied separately in the litterature: intrinsically motivated learning of sensorimotor skills, and socially guided learning of motor skills. This was achieved through the SGIM (Socially Guided Intrinsic Motivation) algorithm. An article presenting this algorithm and associated results obtained the second best student paper award in the IEEE ICDL/Epirob conference, Frankfurt Germany  .
The results of a major large-scale experiment on human-robot interfaces was published in one of the premier venue for human-robot interaction research, the ACM/IEEE HRI conference  . The goal of the experiment, which was very successful and performed out of the lab in Cap Sciences museum in Bordeaux, was to show that the design of adequate interfaces for having a non-engineer human teach new visually grounded words to a robot can improve the performances of learning significantly more than the standard improvement provided by using a sophisticated statistical inference or computer vision algorithms. This is explained by the fact that adequate interfaces allow the robot to collect training examples of a much higher quality.
The FLOWERS team, in collaboration with University Bordeaux I/Labri, has participated as a central actor of the exhibition “Mathematics: A Beautiful Elsewhere” at Fondation Cartier pour l'Art Contemporain in Paris, starting from 19th october 2011 and to be held until 18th march 2012. This installation, called “Ergo-Robots/FLOWERS Fields” was made in collaboration with artist David Lynch and mathematician Mikhail Gromov (IHES, France), and shows computational models of curiosity-driven learning, human-robot interaction as well as self-organization of linguistic conventions. This exhibition, at the crossroads of science and art, has the goal of showing to the general public (several hundred thousands visitors) the intellectual universe of some of the greatest mathematicians of our time as well as some experiments and techniques in other fields (e.g. physics, computer science, robotics) which are directly related to their scientific work. Through this exhibition, the work of the team was also widely disseminated to the general public through large audience radios, magazines and newspapers (France Inter, France Culture, RFI, Sciences et Avenir, Tangente, Financial Times, Daily Telegraph, Libération, ...). More information available at: http://flowers.inria.fr/ergo-robots.php and http://fondation.cartier.com/ .
The Acroban humanoid robot, developped in collaboration with University Bordeaux I/Labri, and which features groundbreaking technologies in terms of flexible morphology (including vertebral column) and capacity for intuitive, safe and robust physical interaction with humans, was highly sollicited for live demonstrations in international venues, both academic and targeted to a wider audience. In particular, on the academic side, Acroban was demonstrated live as a focus demonstration at IEEE IROS 2011 conference in San Francisco, FET 2011 European conference in Budapest, SAME conference in Nice Sophia Antipolis, Foundation of Digital Games 2011 conference in Bordeaux. On the general public side, Acroban was demonstrated at the INNOROBO International Robot Summit in Lyon, at the Laval Virtual salon in Laval, at Fète de la Science in Cité des Sciences et de l'Industrie, during the conference “Des robots et des hommes” at Cité des Sciences et de l'Industrie in Paris. A movie for scientific mediation, explaining the science and technology related to Acroban, was made with the INRIA movie department. Those technologies were also presented and explained on various large audience national and international tv programs (CNBC, TF1, BFM TV, Euronews), radios (France Info, RFI), newspapers and magazines (Le Monde, Le Point, Le Figaro, Le Figaro Magazine, Les Echos, 20 minutes). Finally, Acroban was selected by SmartPlanet.com as one of the best robots of the year 2001 (http://www.smartplanet.fr/smart-technology/compil-2011-un-defile-de-robots-9384/ ).
Best Paper Award : Bootstrapping Intrinsically Motivated Learning with Human Demonstrations in IEEE International Conference on Development and Learning.
S. M. Nguyen, A. Baranes, P.-Y. Oudeyer.